This release is a pre-release and may not be stable for production use.
Agent Framework and ChatKit Integration
This package provides an integration layer between Microsoft Agent Framework and OpenAI ChatKit (Python). Specifically, it mirrors the Agent SDK integration, and provides the following helpers:
stream_agent_response: A helper to convert a streamedAgentResponseUpdatefrom a Microsoft Agent Framework agent that implementsSupportsAgentRunto ChatKit events.ThreadItemConverter: A extendable helper class to convert ChatKit thread items toMessageobjects that can be consumed by an Agent Framework agent.simple_to_agent_input: A helper function that uses the default implementation ofThreadItemConverterto convert a ChatKit thread to a list ofMessage, useful for getting started quickly.
Installation
pip install agent-framework-chatkit --pre
This will install agent-framework-core and openai-chatkit as dependencies.
Requirements and Limitations
Frontend Requirements
The ChatKit integration requires the OpenAI ChatKit frontend library, which has the following requirements:
-
Internet Connectivity Required: The ChatKit UI is loaded from OpenAI's CDN (
cdn.platform.openai.com). This library cannot be self-hosted or bundled locally. -
External Network Requests: The ChatKit frontend makes requests to:
cdn.platform.openai.com- UI library (required)chatgpt.com/ces/v1/projects/oai/settings- Configurationapi-js.mixpanel.com- Telemetry (metadata only, not user messages)
-
Domain Registration for Production: Production deployments require registering your domain at platform.openai.com and configuring a domain key.
Air-Gapped / Regulated Environments
The ChatKit frontend is not suitable for air-gapped or highly-regulated environments where outbound connections to OpenAI domains are restricted.
What IS self-hostable:
- The backend components (
chatkit-python,agent-framework-chatkit) are fully open source and have no external dependencies
What is NOT self-hostable:
- The frontend UI (
chatkit.js) requires connectivity to OpenAI's CDN
For environments with network restrictions, consider building a custom frontend that consumes the ChatKit server protocol, or using alternative UI libraries like ai-sdk.
See openai/chatkit-js#57 for tracking self-hosting feature requests.
Example Usage
Here's a minimal example showing how to integrate Agent Framework with ChatKit:
from collections.abc import AsyncIterator
from typing import Any
from azure.identity import AzureCliCredential
from fastapi import FastAPI, Request
from fastapi.responses import Response, StreamingResponse
from agent_framework import Agent
from agent_framework.openai import OpenAIChatCompletionClient
from agent_framework.chatkit import simple_to_agent_input, stream_agent_response
from chatkit.server import ChatKitServer
from chatkit.types import ThreadMetadata, UserMessageItem, ThreadStreamEvent
# You'll need to implement a Store - see the sample for a SQLiteStore implementation
from your_store import YourStore # type: ignore[import-not-found] # Replace with your Store implementation
# Define your agent with tools
agent = Agent(
client=OpenAIChatCompletionClient(credential=AzureCliCredential()),
instructions="You are a helpful assistant.",
tools=[], # Add your tools here
)
# Create a ChatKit server that uses your agent
class MyChatKitServer(ChatKitServer[dict[str, Any]]):
async def respond(
self,
thread: ThreadMetadata,
input_user_message: UserMessageItem | None,
context: dict[str, Any],
) -> AsyncIterator[ThreadStreamEvent]:
if input_user_message is None:
return
# Load full thread history to maintain conversation context
thread_items_page = await self.store.load_thread_items(
thread_id=thread.id,
after=None,
limit=1000,
order="asc",
context=context,
)
# Convert all ChatKit messages to Agent Framework format
agent_messages = await simple_to_agent_input(thread_items_page.data)
# Run the agent and stream responses
response_stream = agent.run(agent_messages, stream=True)
# Convert agent responses back to ChatKit events
async for event in stream_agent_response(response_stream, thread.id):
yield event
# Set up FastAPI endpoint
app = FastAPI()
chatkit_server = MyChatKitServer(YourStore()) # type: ignore[misc]
@app.post("/chatkit")
async def chatkit_endpoint(request: Request):
result = await chatkit_server.process(await request.body(), {"request": request})
if hasattr(result, '__aiter__'): # Streaming
return StreamingResponse(result, media_type="text/event-stream") # type: ignore[arg-type]
else: # Non-streaming
return Response(content=result.json, media_type="application/json") # type: ignore[union-attr]
For a complete end-to-end example with a full frontend, see the weather agent sample.
5 maintainers
Metadata
5 maintainers
Release files for agent-framework-chatkit 1.0.0b260514
For a detailed explanation of source distributions (sdists) and built distributions (wheels), please see the package formats documentation.
Source distribution (sdist)
| File | Size | Uploaded | |
|---|---|---|---|
| agent_framework_chatkit-1.0.0b260514.tar.gz | 12.7 kB | Details |
Built distribution (wheel)
| File | Interpreter | ABI | Platform | Reset |
|---|---|---|---|---|
| agent_framework_chatkit-1.0.0b260514-py3-none-any.whl | Python 3 | none | any | Details |
Total release size: 24.4 kB
Release files / agent_framework_chatkit-1.0.0b260514.tar.gz
| Download URL | agent_framework_chatkit-1.0.0b260514.tar.gz |
|---|---|
| Size | 12.7 kB |
| Tags | Source |
|
SHA-256 checksum How to use checksums |
fc7e9fd1589c6d8b9f306d8369a98a17b4048ae16c73fe8434b44f5c92fd5687
|
|
BLAKE2b-256 checksum How to use checksums |
691a0a62555c0ea0ceedb2c6bc4fd0ce1f23ff87452f3589fa8d360cdd857c19
|
| Upload date | |
|
Uploaded using Trusted Publishing? What is trusted publishing? |
No |
| Uploaded via |
uv/0.11.6 {"installer":{"name":"uv","version":"0.11.6","subcommand":["publish"]},"python":null,"implementation":{"name":null,"version":null},"distro":null,"system":{"name":null,"release":null},"cpu":null,"openssl_version":null,"setuptools_version":null,"rustc_version":null,"ci":null}
|
Release files / agent_framework_chatkit-1.0.0b260514-py3-none-any.whl
| Download URL | agent_framework_chatkit-1.0.0b260514-py3-none-any.whl |
|---|---|
| Size | 11.8 kB |
| Tags | Python 3 |
|
SHA-256 checksum How to use checksums |
65c334da77d1705689d888c57dfc04faf31f26373bc62d3fa7974f5583517bab
|
|
BLAKE2b-256 checksum How to use checksums |
ad3b0ceb7506f137c9fdd2bb3f246a8b9b58829c963cf2bef63b880760d9606b
|
| Upload date | |
|
Uploaded using Trusted Publishing? What is trusted publishing? |
No |
| Uploaded via |
uv/0.11.6 {"installer":{"name":"uv","version":"0.11.6","subcommand":["publish"]},"python":null,"implementation":{"name":null,"version":null},"distro":null,"system":{"name":null,"release":null},"cpu":null,"openssl_version":null,"setuptools_version":null,"rustc_version":null,"ci":null}
|
